Table 4.

Whole exome sequence filtering data for two UDP families

Family	Operation	Cumulative Residual Variants"
UDP 0337	None	120,469
	1000 genome 1%	59,482
	SNP chip based family linkage filec	53,087
	Kill filed	50,467
	Homozygous recessive	11
	Homozygous recessive + not in db130e	3f
	Compound heterozygousg	121
	Compound heterozygous + not in db130	71
UDP 0984	None	135,292
	1000 genome 1%	68,200
	SNP chip based family linkage filec	52,426
	Kill filed	48,532
	Homozygous recessive	39
	Homozygous recessive + not in db130c	8
	Compound heterozygousf	286
	Compound heterozygous + not in db130	143
UDP1706	None	122,195
	1000 genome 1%	52,683
	SNP chip based family linkage filed	28,359
	Kill filee	28,359
	Dominant	1,675
	Dominant + not in db130	872
	Dominant + not in db130 + deleterious coding	64

^aThese lists can be further ranked or sorted by several means, e.g., by predicted pathogenicity.

^bMutation types include frameshifting indels, non-synonymous missense mutations, canonical splice site mutations, stop mutations.

^cThe SNP chip based family linkage file can have a much larger impact than seen for these two families, depending on where recombinations occurred and how many participants with informative meioses are available.

^dThe Kill file includes genes in the HLA region, very highly polymorphic genes such as the mucin genes, genes that code for the taste and olfactory receptors, and known pseudogenes with open coding regions.

^eUsing the “not in db130 filter” needs to be done with caution because it contains pathogenic variants.

^fIncludes AFG3L2.

^gFor compound heterozygotes, the filter includes any combination of variants that follow a predicted Mendelian pattern. In practice, however, one of the two variants identified will be intronic, 3' UTR, etc. Therefore, one can look at only the genes with two “good” mutations. For example, in for UDP 0984, the list of 143 variants only represents 17 genes with two mutations of the type surveyed. One of those 17 was the responsible gene, GLB1.